摘要
针对TV模型存在分块效应,而四阶PDE模型具有保持平坦区域光滑性的特点,提出自适应耦合TV和四阶PDE的正则化图像放大模型。根据图像内容合理调整耦合系数,在图像渐变和平坦区域运用四阶PDE扩散,消除分块效应;而在图像的突变区域运用TV模型滤波,保持突变边缘。实验结果表明,该算法是一种有效的图像放大方法。
A regularized image zooming model of adaptive couple TV and fourth-order PDE is proposed according to that the fourth-order PDE model can maintain the smoothness of flat areas and the total variation model has blocky effects.Based on the above considerations,the coupling coefficients were reasonably adjusted according to area contents.In the flat areas,the fourth-order PDE model was emphasized to eliminate block effects;where as in fast changing region,TV model was stressed to enhance the large gradient edges.Experimental results show that the algorithm is an effective image enlargement method.
出处
《现代电子技术》
2011年第9期80-82,86,共4页
Modern Electronics Technique
基金
国家自然科学基金资助项目(60873196)
关键词
图像放大
偏微分方程
总变分
高阶PDE
image zooming
partial differential equations(PDE)
total variation
higher-order PDE